Hi Laura We have a 4-input (2 dual pole) 512 channel correlator working in a ROACH working at 200MHz (sampling at 800MHz). Each of the 4 input channels has the following; - coarse delay - 1k PFB with 4 taps PFB FIR - fine delay and fringe-stopping block - re-quantisation to 4 bits - multiplication to produce the products - accumulation in DRAM
The FPGA is full to 98% (logic). I am unsure if this relates to your design. It is in the pocket correlator folder as poco_roach. Regards Andrew On 14 May 2010 05:24, Laura Spitler <[email protected]> wrote: > Hi Dan, > > The FPGA is clocked at 200 MHz. I used the 3 Gs/s ADC yellow block > running at 1600 MHz feeding a 8 input wideband FFT. The timing errors > were in the FFT. > > Laura > > > > On Thu, May 13, 2010 at 10:58 PM, Dan Werthimer <[email protected]> > wrote: > > > > hi laura, > > > > are you trying to clock your FGPA at 200 or 400 MHz? > > > > if you configure the ADC yellow block to input 4 samples per clock, > > and sampling at 1600 Msps, then you will need to clock your > > fpga at 400 MHz, which is pretty near impossible for virtex 5 > > and you'll get lots of timing errors. > > > > if you configure the yellow block to input 8 samples per clock, > > and sampling at 1600 Msps, then the FGPA is clocking at 200 MHz, > > and routing and timing should work. > > > > best wishes, > > > > dan > > > > On 5/13/2010 6:46 PM, Laura Spitler wrote: > >> > >> Hi everyone, > >> > >> Does anyone have a sense of whether a dual polarization, 800 MHz > >> spectrometer with ~2k channels would fit into a ROACH? I have a test > >> design using the ADC083000 and and two small wideband FFTs. It seems > >> to work in terms of utilization, but I get a crazy number of timing > >> errors. > >> Has anyone successfully built anything like this? > >> > >> Thanks, > >> Laura > >> > >> > > > > > > > >
